1. Riker, William, The Theory of Political Coalitions (New Haven: Yale University Press, 1962), p. 8Google Scholar.

2. Moon, J. Donald, “The Logic of Political Inquiry: A Synthesis of Opposed Perspec-tives,” in Handbook of Political Science, ed. Greenstein, Fred I. and Polsby, Nelson W. (Reading, Mass.: Addison Wesley, 1975), vol. 1, pp. 130–206, esp. p. 176Google Scholar.

3. For critiques of causal social science, see Maclntyre, Alasdair, “How Is a Comparative Science of Politics Possible?” in Maclntyre, , Against the Self-Images of the Age (New York: Schocken, 1971)Google ScholarTaylor, Charles, Philosophy and the Human Sciences (Cambridge: Cam-bridge University Press, 1985), p. 55. For critiques of historical/interpretive work, see Riker, Political Coalitions, p. viii, and subsequent writings, as well asGoogle ScholarChubb, John E. and Moe, Terry M., “Controversy: Should Market Forces Control Educational Decision Making?” American Political Science Review 84 (1990): 558–67. For arguments that the approaches can coexist peacefully without necessarily being integrated.Google ScholarTulis, Jeffrey K., The Rhetorical Presidency (Princeton: Princeton University Press, 1987), p. 14Google ScholarHurley, Susan L., Natural Reasons: Personality and Polity (New York: Oxford University Press, 1989), p. 98 (the two modes of explanation “do not compete with one another, nor is there any incompatibility between them”). For arguments of their inseparability, see Moon, “Political Inquiry”Google ScholarJohn, Fere-john, “Rationality and Interpretation: Parliamentary Elections in Early Stuart England,” in The Economic Approach to Politics: A Critical Reassessment of the Theory of Rational Action, ed. Monroe, Kristen Renwick (New York: HarperCollins, 1991), pp. 279–305Google ScholarDavidson, Donald, Essays on Actions and Events (New York: Oxford University Press, 1990), esp. pp. 207–44, discussed briefly belowGoogle Scholar.

4. There are currently “new institutionalisms” throughout the social sciences, including economics, history, organization theory, and sociology, but they tend to align with these two general perspectives in political science. The organizational economists and economic historians, for example, work in ways generally consonant with rational choice “new institutionalists” studying politics. Social historians, organizational theorists, and sociologists find more in common with political scientists in the historical/interpretive “new institutionalise” camp. For a lucid analysis of these trends that introduces an important collection of “new institutionalist” essays in sociological organizational theory. DiMaggio, Paul J. and Powell, Walter W., “Introduction,” in The New Institutionalism in Organizational Analysis, ed. Powell, and DiMaggio, (Chicago: University of Chicago Press, 1991), pp. 1–38Google Scholar.

5. March, James G. and Olsen, johan P., Rediscovering Institutions: The Organizational Basis of Politics (New York: Free Press, 1989), pp. 1–52Google Scholar.

6. Chubb and Moe, “Controversy,” p. 565.

7. Matthew D. McCubbins, Roger G. Noll, and Barry R. Weingast, “Positive and Normative Models of Due Process: An Integrative Approach to Administrative Procedures” (Hoover Institution Working Paper P-90-10, Stanford, 1990); and Ferejohn, “Rationality and Interpretation.”

8. I use “determinism” here to refer to accounts that explain the behavior of political actors entirely in terms of factors outside the actors' conscious control. I therefore do not mean that all causal accounts are deterministic. For example, propositions that agents' actions are affected though not wholly determined by external physical causes, that their acltions may be caused by what they take to be good reasons, and that their acts are themselves physical causes of other events are all not deterministic in this sense.

9. Such claims are currently most prevalent among rational choice theorists, but others concur. For example, a Marxist scholar, Paul Cammack, has argued similarly against Theda Skocpol's historical and institutional sociology. It leads, Cammack says, to abandonment “of any kind of coherent theoretical framework at all, in favour of middle-level enquiries into a multiplicity of issues in a multiplicity of settings in the hope that something will turn up. Studies pitched in the middle of nowhere are not likely to lead anywhere” (Cammack, Paul, “Review Article: Bringing the State Back In?” British Journal of Political Science 19 [1989]: 261–90, at p. 287). Here I suggest instead that studies pitched, not nowhere, but in a number of theoretical places all trying to account for common experiences, are more likely to be fruitful than staking all on one grand theory in advanceCrossRefGoogle Scholar.

10. “Agency” as used here is to be understood not in the legal or economic sense, in which one person is an “agent” working on behalf of another, but rather in the philosophical sense, in which “agents” are viewed as causally potent, morally responsible actors guided by their own intentions, not external forces. Following much contemporary literature, I also use “agents” and “actors” interchangeably. Both terms are in fact rather unsatisfactory because they are both ambiguous on the fundamental issue of whether a person'conduct should be attributed to that person or to some other, a “principal” or “author.” Unfortunately, no better substitute has yet suggested itself.

11. Charles Taylor has drawn a similar and quite influential distinction between views of agency centered on strategic planning and ones stressing “strong evaluation” of the ends of agents {Human Sciences, pp. 102–5). For Taylor, however, judgments of ends seem to be more matters of genuine discovery of the purposes with which we are socially (and perhaps naturally) endowed, rather than reasoned but alterable choices among these purposes, as I hold here. This point is discussed further below.

12. See Dunn, John, “Understanding Revolutions,” in Dunn, Rethinking Modern Political Theory: Essays 1979–84 (Cambridge: Cambridge University Press, 1985), pp. 74–78Google ScholarBetts, Katharine, “The Conditions of Action, Power and the Problem of Interests,” Sociological Review 34 (1986): 39–64Google Scholar.

13. Anspach, Renee R., “From Stigma to Identity Politics: Political Activism Among the Physically Disabled and Former Mental Patients,” Social Science and Medicine 13 (1979): 765–73Google Scholar.

14. Smith, Rogers M., Liberalism and American Constitutional Law, pap, . ed. (Cambridge: Harvard University Press, 1990). For the leading contemporary theoretical statement of this type of liberalism (with which I nonetheless have disagreements)Google ScholarGalston, William A., Liberal Purposes: Goods, Virtues and Diversity in the Liberal Slate (Cambridge: Cambridge University Press, 1991)Google Scholar.

15. Aristotle, , Nicomachaean Ethics, trans. Ostwald, M. (Indianapolis: Bobbs-Merrill, 1962), p 1094aGoogle Scholar.

16. Hobbes, Thomas, Leviathan (Great Britain: Penguin, 1971), p. 728Google Scholar.

17. Seidelman, Raymond and Harpham, Edward J., Disenchanted Realists: Political Science and the American Crisis, 1884–1984 (Albany: State University of New York Press, 1985)Google Scholar.

18. Gabriel A. Almond, “Approaches to Developmental Causation,” in Crisis, Choice, and Change: Historical Studies of Political Development, ed. Gabriel A. Almond, Scott C. Flanagan, and Robert J. Mundt (Boston: Little, Brown, pp. 1–42, esp. pp. 6, 12–13; and Almond, Gabriel A. with Genco, Stephen, “Clouds, Clocks, and the Study of Politics,” in Almond, A Discipline Divided: Schools and Sects in Political Science (Newbury Park, Calif.: Sage, 1990), pp. 32–65Google Scholar (arguing at page 45 that the “elements of the implicit logic that informs much of political science research today appears to imply a substantive model of the political world which closely resembles the deterministic ‘clock model’ outlined by Popper”).

19. Stephen D. Krasner, “Global Communication and National Power: Battle of the Sexes versus Prisoner's Dilemma or, Life on the Pareto Frontier” (Paper delivered to the Law and International Politics Seminar, Yale Law School, New Haven, March 25, 1991), p. 37.

20. For example, in 1973 Gabriel Almond was still saying that if structural functionalism “falls short of explanation in the tight sense of experimental science then it suffers the difficulties encountered by all science ‘that nature is capricious at the quantum level’ and that ‘the order of systems emerges as a statistical reality out of the disorder of particles’ ” (Almond, “Approaches,” p. 6). The remainder of that excellent essay, focused on the importance of human political choices, did not really treat them as so random as this statement implies; and Almond explicitly disavowed “the quantum-jump model” as a mischaracterization of “rational human behavior” in Almond and Genco (“Clocks,” p. 34).

21. Again using Almond's discussions in his 1973 coedited volume as an example, he there admitted that he and his coworkers had dealt with the political choices of leaders only “impressionistically.” Thus “leadership phenomena tend to be treated residually” in those studies, even though Almond believed leadership choices to be “the sufficient explanation of a specific outcome” in most cases (Gabriel A. Almond and Robert J. Mundt, “Crisis, Choice, and Change: Some Tentative Conclusions,” in Almond et al., Crisis, pp. 619–49, esp. pp. 621, 648–49).

22. Almond and Genco similarly argue that although political scientists do not “actually see” the political world deterministically and often give more complex pictures in their detailed empirical analyses, the “meta-methodological principles and procedures” they have “borrowed from the physical sciences—or more correctly, from a certain philosophical perspective on the physical sciences” have come to the discipline “with an array of substantive assumptions,” committing it to a “research program designed to strip away the … purposive aspects of political reality in order to expose its ‘true’ clocklike structure” (“Clocks,” p. 45).

23. Almond, and Genco, , “Clocks, p. 36; and Ricci, David M., The Tragedy of Political Science (New Haven: Yale University Press, 1984), pp. 250–58Google Scholar.

24. Tsebelis, George, “Controversy, Crime and Punishment: Are One-Shot, Two-Person Games Enough?” American Political Science Review 84 (1990): 576‐86Google Scholar, at p. 579; and Ferejohn, “Rationality and Interpretation,” pp. 282, 284–85.

25. Several readers have suggested that prediction of complex major events is too demanding a test, noting that seismologists did not predict the 1989 San Francisco earthquake. But they did predict major earthquakes at some point in the future, and they can provide persuasive retrospective accounts of past earthquakes. Most modern political science (including Marxian analyses) suggested instead that resurgent fundamentalism on a large scale, as in Iran and elsewhere, was quite unlikely. We are, moreover, far from achieving generally persuasive “postdictive” accounts of these phenomena. I do not take these facts as proof that we will never reach a more mature and powerful ”science” of politics, only as evidence that we are, at best, still far from such a thing.

26. March, James G. and Olsen, Johan P., “The New Institutionalism: Organizational Factors in Political Life,” American Political Science Review 78 (1984): 734–49Google Scholar, esp. pp. 735–38; March and Olsen, Rediscovering, p. 6; Wendt, Alexander E., “The Agent-Structure Problem in International Relations Theory,” International Organizations 41 (1987): 335–70Google Scholar, esp. p. 356; and Smith, Rogers M., “Political Jurisprudence, the ‘New Institutionalism,’ and the Future of Public Law,” American Political Science Review 82 (1988): 89–108Google Scholar, esp. pp. 94, 10.

27. For relevant discussions, see Takaki, Ronald, Iron Cages: Race and Culture in Nineteenth Century America (New York: Oxford University Press, 1979)Google Scholar; and Gordon, Robert W., “Critical Legal Histories,” Stanford Law Review 36 (1984): 57–125Google Scholar. Charles Taylor has made a similarcriticism of followers of Derrida, whose emphasis on the constitutive role of self-referential languages can render “the code as ultimate, dominating the supposedly autonomous agent” (Taylor, Human Sciences, p. 11). Many “new institutionalists” in sociology, such as John W. Meyer et al., argue explicitly that they see “action as the enactment of broad institutional scripts rather than a matter of internally generated and autonomous choice, motivation and purpose” (Meyer, John W.Boli, John, and Thomas, George M., “Ontology and Rationalization in the Western Account, ” in Institutional Structure: Constituting State, Society, and the Individual, ed. Boli, Meyer, and Thomas, ([Newbury Park, Calif.: Sage, 1987], pp. 12–37Google Scholar, esp. p. 13). I think this overstates the determining powers of such “scripts.”

28. Smith, “New Non-Science.”

29. See, e.g., Shepsle, Kenneth A., “Institutional Arrangements and Equilibrium in Multi-Dimensional Voting Models,” American Journal of Political Science 23 (1979): 27–59Google ScholarShepsle, , “Institutional Equilibrium and Equilibrium Institutions,” in Political Science: The Science of Politics, ed. Weisberg, Herbert F. (New York: Agathon, 1986), pp. 51–81Google ScholarShepsle, Kenneth A. and Weingast, Barry R., “Structure-Induced Equilibrium and Legislative Choice,” Public Choice 37 (1981): 503–19Google ScholarShepsle, and Weingast, , “When Do Rules of Procedure Matter?” Journal of Politics 46 (1984): 206‐21Google Scholar; and McCubbins, Matthew D.Noll, Roger G., and Weingast, Barry R., “Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies,” Virginia Law Review 75 (1989): 431–82Google Scholar. Writing less formally, Terry Moe has evaluated and contributed to this body of work in a number of significant essays. See, e.g., Moe, Terry M., “The Politics of Bureaucratic Structure,” in Can the Government Govern? ed. Chubb, John E. and Peterson, Paul E. (Washington, D. C.: Brookings Institution, 1989), pp. 267–327;Google Scholar“The Politics of Structural Choice: Toward a Theory of Public Bureaucracy,” in Organization Theory: From Chester Barnard to the Present and Beyond, ed. Williamson, Oliver E. (New York: Oxford University Press, 1990):Google Scholar and “Political Institutions: The Neglected Side of the Story” (Paper presented to a conference on “The Organization of Political Institutions,” Yale Law School, New Haven, April 27–28, 1990), and references therein.

30. Riker, William H., “Implications from the Disequilibrium of Majority Rule for the Study of Institutions, ” American Political Science Review 74 (1980): 432–46Google Scholar; Riker, , The Art of Political Manipulation (New Haven: Yale University Press, 1986)Google Scholar; Cox, Gary W., The Efficient Secret: The Cabinet and the Development of Political Parties in Victorian England (Cambridge: Cam-bridge University Press, 1987)Google Scholar; Shepsle, Kenneth A. and Weingast, Barry R., “The Institutional Foundations of Committee Power,” American Political Science Review 81 (1987): 85–104Google Scholar; and Grofman, Bernard and Wittman, Donald, eds., The Federalist Papers and the New Institutionalism (New York: Agathon, 1989)Google Scholar.

31. Shepsle, “Institutional Equilibrium,” p. 51.

32. Riker, Political Manipulation, pp. ix–x.

33. Jensen, Michael C. and Meckling, William H., “Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure,” Journal of Financial Economics 3 (1976): 305–60Google Scholar; Fama, Eugene C. and Jensen, Michael C., “Agency Problems and Residual Claims,” Journal of Law and Economics 26 (1983): 327–49Google Scholar, esp. p. 329; and discussed in Barzelay, Michael and Smith, Rogers M., “The One Best System? A Political Analysis of Neoclassical Institutionalist Perspectives on the Modern Corporation,” in Corporations and Society: Power and Responsibility, ed. Samuels, Warren J. and Mille, Arthur S. (Westport, Conn.: Greenwood, 1987), pp. 81–110Google Scholar.

34. Roger G. Noll and Barry R. Weingast, “Rational Actor Theory, Social Norms, and Policy Implementation: Applications to Administrative Processes and Bureaucratic Culture,” in Monroe, Economic Approach, pp. 239–43.

35. In this respect rational choice has a heavier burden of proof. If rational choice models are to be plausible guides to political conduct, they must approximate reasonably well the behavior of most political actors most of the time; the odd moment of rationality in a nonrational world would probably not be so significant as to make it the center of analysis. But even if transformations of basic interests and identities via normative persuasion or deliberation are quite infrequent, they are so fundamental and constitutive for “normal” politics that they must be prominent in political science.

36. Let me emphasize that I do not assume these ways are anything close to infinite in number. People are, of course, socially as well as physically constituted through ongoing interactions with their various environments. Those forms of constitution set limits on what they can plausibly be, indeed on what they hope to be. But as argued more fully below, people are endowed with capabilities in ways that empower as well as constrain them. Usually these capabilities enable them to decide on more than one sense of who they are and what they wish to do and become.

37. As noted above, Davidson has offered perhaps the most influential philosophical attempt to reconcile the physical causality of the natural sciences with the causality via reason and ideas that seems necessary for intentional human agency. My only objection to his view is that it is sometimes taken to mean that we no longer have to be concerned about the intersection of “physical” with “ideational” or “intentional” causality. Analyses can discuss the same phenomena within essentially separate frameworks while blithely asserting their compatibility. I think instead that if we are committed to seeing our existence whole, then the relationship of “blind” physical mechanisms to “intentional” causality via reasons must remain on the agenda of scientific inquiry. Most such inquiry will, to be sure, be outside the practical scope of most political science, but that does not mean political scientists can be blind to pertinent developments (cf. Galston, Liberal Purposes, p. 37). Davidson can plausibly be read in ways that avoid this concern. Although in principle physical and ideational causality, if they are indeed both true, must somehow dovetail in explicable ways, it is questionable whether on its own terms either sort of account can ever produce fully determinate explanations of human conduct. It is in particular not at all clear that physical accounts will often be able to render more determinate predictions than intentionalist ones. Yet conduct is somehow in the end determined. It may be that in many instances the indeterminacies that would remain even in the best conceivable physical expla-nations of human brain processes are great enough so that the most powerful explanation of how those indeterminacies are actually resolved will be provided by interpretations of the agent's mental reasoning. In those cases we could assert that an intentionalist account is more illuminating than a physical or biological one, even though the latter is not wrong. But this is a claim that should be recurringly earned by comparing the power of ever-improving physical explanations with intentionalist ones. It cannot simply be assumed to be so. I am grateful to John Dunn for pressing these issues.

38. Dunn, John, Modem Revolutions: An Introduction to the Analysis of a Political Phenomenon, 2d ed. (Cambridge: Cambridge University Press, 1989), p. 16CrossRefGoogle Scholar.

39. Moe, , “Bureaucratic Structure”; and Katzmann, Robert A., Institutional Disability: The Saga of Transportation Policy for the Disabled (Washington, D. C.: Brookings Institution, 1986). The elements of Katzmann's account stressed here are affirmed in all the discussions of disability politics I have found. See especiallyGoogle ScholarScotch, Richard K., From Good Will to Civil Rights: Transforming Federal Disability Policy (Philadelphia: Temple University Press, 1984), and also Anspach, “Identity Politics”Google ScholarEisenberg, Myron G.Griggins, Cynthia and Duval, Richard J., “Introduction,” in Disabled People as Second-Class Citizens, ed. Griggins, Eisenberg and Duval, , (New York: Springer, 1982), pp. 13–19Google Scholar; Hahn, Harlan, “Introduction: Disability Policy and the Problem of Discrimination,” American Behavioral Scientist 28 (1985): 293–318Google Scholar; Berkowitz, Edward D., Disabled Policy: America's Programs for the Handicapped (New York: Cambridge University Press, 1987)Google ScholarDriedger, Diane, The Last Civil Rights Movement: Disabled Peoples' International (London: Hurst, 1989)Google Scholar.

40. Moe, “Bureaucratic Structure,” p. 268; “Structural Choice,” p. 117; and “Political Institutions,” p. 60.

41. Moe, “Bureaucratic Politics,” pp. 269, 281–84; and “Structural Choice,” p. 129.

42. Moe, “Bureaucratic Politics,” pp. 289–92, 297–98, 309–10, 316, 323, 328.

43. Katzmann, Institutional Disability, pp. 9–14, 25, 76–77, 80–83, 189 n. 1. Perhaps his fullest methodological self-description is on page 81, where he links his work to James Q, Wilson's: “My discussion of the administrative process will show that no single approach can explain the administrative course over time, that its twists and turns are the product of the interplay of various factors, and that the primacy of one or another may depend on particular circumstances and shifting constraints.” This is a virtual denial of the possibility of general theory for administrative processes.

44. Richard Scotch notes that up to the 1960s most groups representing the disabled simply sought “expanded benefits” for their particular clienteles. Only the National Federation of the Blind displayed a “strong civil rights orientation,” and “none was oriented toward the general issue of civil rights for all disabled people” (Good Will, pp. 33–34, 57). Renee Anspach similarly holds that, until the 1960s, “there were simply no alternatives to the ideology of rehabilitation” (“Identity Politics,” p. 771). Myron Eisenberg et al. agree that the disabled long accepted “the treatment/rehabilitation/sympathy model” and “the role it prescribed for them–patients, whether in or out of the hospital, and second-class citizens” (Second-Class Citizens, p. xiv). Harlan Hahn reports that some academic researchers had long perceived the disabled as a minority group denied civil rights, but says this view was “tentative” and “secondary or peripheral.” As the disability civil rights movement emerged, many more of the disabled came to see themselves as members of such a minority (“Introduction,” pp. 300–301).

45. I cannot explore the sources of these victories here, but these positions clearly gained significant political support from Americans' experiences during first the Depression and then large-scale warfare, hot and cold, against enemies with anti-rights ideologies. Thus the circumstances in which the disabled acted should be seen as products of politics, not in any way exogenous to it.

46. Katzmann, Institutional Disability, pp. 20–31, 36–40, 45–49, 99. See also Scotch, Good Will, pp. 40, 51–59, 141, 148–49; Hahn, “Introduction,” p. 299; and Berkowitz, Disabled Policy, pp. 212–13.

47. Katzmann, Institutional Disability, pp. 7, 38–40, 55, 58, 61–66, 76, 109–13.

48. Katzmann, Institutional Disability, pp. 99, 101, 110–11, 159–60. For similar accounts, indicating that many disabled activists were in fact veterans of civil rights and feminist campaigns, see Anspach, “Identity Politics,” pp. 765–73; Eisenberg et al., Second-Class Citizens, pp. xv–xvi; Scotch, Good Will, pp. 6, 35; Hahn, “Introduction,” p. 300; Berkowitz, Disabled Policy, pp. 189, 197, 200, 229; and Driedger, Last Civil Rights Movement, pp. 11–12. The identification with the civil rights movement went so far that disability organizations protesting against delay in enforcing Section 504 occupied H. E. W. Secretary Joseph Califano's office and sang “We Want 504” to the tune of “We Shall Overcome.” Anspach notes that the new disability activist groups “consciously” endeavored “to alter both the self-concepts and societal conceptions of their participants” (“Identity Politics, ” p. 766).

49. Katzmann, , Institutional Disability, pp. 3, 7–8, 27, 30, 114, 155, 189. In 07, 1991Google Scholar, Justin Dart, Jr., head of the President's Commission on the Employment of People with Disabilities, and his executive assistant John Lancaster were both denied access to a relatively small commercial airplane because they were in wheelchairs. The denial was consistent with D. O. T. regulations, but Lancaster said, “It made me feel like a second-class citizen” (Holmes, Stephen A., “Airline Denies Seat to Disabled Official,” New York Times 07 16, 1991, p. A10)Google Scholar.

50. Katzmann, Institutional Disability, pp. 50–53, 60–63, 75–79, 90–132, 190–92; and Moe, “Bureaucratic Structure,” pp. 274–76, and “Structural Choice,” pp. 136–37.

51. The arbitrariness of choosing to analyze only certain kinds of time periods, those in which preferences are relatively fixed, is especially striking in analyses that emphasize, as Moe repeatedly does, that the “game of structural politics never ends” (Moe, “Bureaucratic Structure,” p. 284, and “Structural Choice,” p. 146).

52. Scotch similarly argues that although Section 504 probably could not “have been enacted prior to the late 1960s,” it nonetheless represented a “creative fusion” of rehabilitation and civil rights concepts, not “an inevitable byproduct” of that era. He also observes that even though disability groups did to some degree develop a “party line” in favor of a civil rights approach after 504 was enacted, nonetheless different “segments of the movement have emphasized different and sometimes conflicting goals” (Good Will, pp. 101, 152, 167).

53. For similar assessments of the strategic and intrinsic advantages of the civil rights/ mainstreaming perspective, see Scotch, Good Will, pp. 156–64; Harlan, “Introduction,” pp. 300–301; and Katzmann, Institutional Disability, pp. 189, 202–4.

54. McCubbins, Noll, and Weingast, “Positive and Normative Models,” pp. 1–2, 12–13, 23–24.

55. Smith, “Political Jurisprudence.”

56. For a more successful integration of rational choice and historical analysis by one of these authors, see Barry R. Weingast, “The Political Economy of Slavery: Credible Commitments and the Preservation of the Union, 1800–1860” (Paper presented at the Annual Meeting of the American Political Science Association, Washington, DC, 1991).

57. Ferejohn, “Rationality and Interpretation,” pp. 282–86, 297–99.

58. Ferejohn, “Rationality and Interpretation,” p. 299.

59. Ferejohn, “Rationality and Interpretation,” pp. 285, 298–99.

60. In this regard Ferejohn notes but does not specifically address Charles Taylor's claim that the “most fundamental reason for the impossibility of hard prediction is that man is a self-defining animal” (Taylor, Human Sciences, p. 55, quoted in Ferejohn, “Rationality and Interpretation,” p. 300 n. 4).

61. Maclntyre, “Comparative Science”; Moon, ”Political Inquiry,” pp. 155, 176; and Taylor, Human Sciences, pp. 15–57.

62. Taylor, at least, seems to agree, since he insists repeatedly that the human sciences are indeed sciences, albeit interpretive ones, and that they involve examining how far evidence supports hypotheses (Taylor, Human Sciences, pp. 50, 58, 117–18).

63. Taylor, Human Sciences, pp. 17–18, 53.

64. Rosenstone, Steven J., Forecasting Presidential Elections (New Haven: Yale University Press, 1983)Google Scholar.

65. Such broader “new institutionalise” views in political science include Wendt, “Agent-Structure Problem,” pp. 337–39; Krasner, Stephen D., “Sovereignty: An Institutional Perspective,” Comparative Political Studies 21 (1988): 66–94; March and Olsen, Rediscovering Institutions;Google ScholarKarl, Terry Lynn, “Dilemmas of Democratization in Latin America,” Comparative Politics 23 (1990): 1–21Google Scholar; and Hall, Peter A., “The Movement from Keynesianism to Monetarism: Institutional Analysis and British Economic Policy in the 1970s,” in Historical Institutwnalism: Politics, Society, Economy, ed. Steinmo, SvenThelen, Kathleen and Longstreth, Frank, (New York: Cambridge University Press, forthcoming 1992)Google Scholar. My suggestions also obviously draw on the arguments of many historical sociologists, particularly Skocpol, Theda, “Sociology's Historical Imagination,” in Vision and Method in Historical Sociology, ed. Skocpol, (Cambridge: Cambridge University Press, 1984), pp. 1–21Google Scholar. They may seem especially to resemble the case for “structurational” analyses Anthony Giddens has made in an astonishingly large number of works. See, e.g., Giddens, Anthony, The Constitution o} Society: Outline of the Theory ofStructuration (Berkeley: University of California Press, 1984)Google Scholar. My aim here, however, is to sketch a framework for designing research projects that can enable quite different hypotheses about structure and agency to be explored. Giddens's primary concern is to offer a set of such hypotheses. Although they are certainly ones that should be examined in the ways I am endorsing, I share Michael Taylor's concern that, by heavily stressing the “mutually constitutive” character of structures and agents, Giddens's position makes it hard to draw causal arrows (Taylor, Michael, “Structure, Culture and Action in the Explanation of Social Change,” Politics and Society 17 [1989}: 115Google Scholar–62, esp. pp. 117–18). This difficulty is related to the more normative one discussed in the text, about whether decisions of agents should be seen as “discoveries,” because the ways people are constituted leave them no other genuine choices, or instead as reasoned choices. In those cases where it was instantiated, the latter view would make the attribution of causality to agents themselves more plausible.

66. Some “new institutionalise” writers have begun to offer classifications of different types of basic institutions and the differing capabilities, including capabilities for change, they provide (Krasner, “Sovereignty”; and Hall, “Movement from Keynesianism”). Most nonetheless prefer as I do to describe a research approach that permits all candidates for significant structures and agent capabilities to be explored. Thus Skocpol calls generally for examining “the interplay of meaningful actions and structural contexts” (“Historical Imagination,” p. 1); and Terry Karl calls for “an interactive approach that seeks explicitly to relate structural constraints to the shaping of contingent choice” (“Dilemmas,” p. 1).

67. For similar suggestions, see Almond and Mundt, “Crisis,” p. 637; and especially Karl, “Dilemmas,” pp. 7–8.

68. Krasner, “Sovereignty,” pp. 67, 72, makes these points in just these terms and, like most new institutionalists, stresses the “path dependent” character of the analyses that result. See also Karl, “Dilemmas,” p. 7; and William H. Sewell Jr., “Three Temporalities: Toward a Sociology of the Event” (Comparative Studies of Social Transformations Working Paper 58, Ann Arbor, 1990), p. 16.

69. Karen Orren and Stephen Skowronek, for example, argue powerfully that particular structures may display largely independent paths of development and interact with others in “multilayered” ways that can only be grasped retrospectively, not predicted (Karen Orren and Stephen Skowronek, “Beyond the Iconography of Order: The New Institutionalism in Political Analysis,” Paper presented at the Annual Meeting of the American Political Science Association, Washington, DC, 1991).

70. Or at least in sociology: Anspach, “Identity Politics,” pp. 771-73, and Scotch, Good Will, pp. 152–53, 156–64, offer discussions in this spirit.

71. Sandel, Michael J., Liberalism and the Limits of Justice (New York: Cambridge University Press, 1982), pp. 175–83Google Scholar.

72. Almond and Mundt, “Crisis,” pp. 637–38, describe such an anticipated discovery. The Mexican leader Cardenas fostered a coalition that was outside the set they had thought possible based on existing preferences.

73. Rossiter, Clinton, ed., The Federalist Papers (New York: New American Library, 1961), pp. 81–82Google Scholar.